Protein Crystallography: Automating a Temperamental Science
How does the lab manager best implement automated technologies to optimize experimental workflow for improved cost-and time-efficiency?
With the evolution of proteomics and the completion of the human genome project, protein crystallography has developed into a key area of pharmaceutical research. The sensitivity of this technique for identifying three dimensional protein structures means it has proved invaluable for the process of rational drug design. As a result, the protein crystallography laboratory has been subject to growing pressures. As technological advances have allowed screening laboratories to increase throughput and maximize hit-to-lead success, it was rapidly realized that in order to prevent bottlenecks in the discovery pipeline, the protein crystallography lab must also keep pace. This has not been easy as individual proteins or protein families have specific requirements and crystallography methodologies and manual techniques are traditionally used for crystallization set-up and optimization. Nevertheless, in recent years technology providers have developed innovative technologies for the protein crystallography lab without compromising experimental flexibility and most of all, data quality. But what has this meant for the lab manager? What are the financial implications of such investment? How has the move to automated technologies been accepted by the scientists in what was once a very “hands-on” discipline?
AN INTEGRATED RESEARCH SET-UP
The role of many protein crystallography teams within pharma is predominantly to provide support to lead discovery programs. As such, they are very closely linked with many research areas and departments across the whole enterprise. The crystallography team is primarily involved with the crystallization and characterization of proteins. This includes activities such as screening, optimization of hits, crystal production, diffraction data collection, and structure determination. However, interactions with other teams, both upstream and downstream of the protein crystallography group, are essential for the successful contribution to any drug discovery program. These teams may include molecular biologists, protein engineers, and biophysicists to identify the targets and supply the proteins to be crystallized, in addition to computational and medicinal chemists, cell biologists, and biochemists, for downstream validation and further discovery efforts. This set-up often means that the crystallography team must not only be dynamic and flexible to meet changing demands but also have the combined expertise and crystallography know-how to deliver within a set time frame. Such enterprise-wide activities require seamless or near seamless interactions across departments and multiple sites worldwide, with regular meetings to address issues and reevaluate objectives.
Providing this level of support requires the protein crystallography lab manager to understand the needs of the various therapeutic areas and prioritize targets efficiently. Obviously, one of the associated pressures is effective time management. The timely delivery of crystal structures can greatly assist the decision making process as to what leads to follow and how. Additionally, the protein crystallography laboratory has to remain a viable source of data for the associated discovery efforts.
With the endless number of protein structures that are yet to be determined, it can be tempting for a crystallography team to place their own focus on specific protein groups or therapeutic areas. However, it is critical that pharma teams remain focused on targets specified by their collaborative departmental groups. Individual companies may also have a particular disease or therapeutic focus for which specific targets would be more relevant and have greater impact commercially. Such priorities mean additional pressure for protein crystallography groups as they may necessitate focusing on proteins that are more difficult to crystallize and require more time and effort to establish a successful and reliable crystallography method.
THE IMPORTANCE OF TEAMWORK
To meet these demands, one challenge for the pharma lab manager is to strategically distribute crystallography programs between team members taking into consideration the expertise and available time of members, and the equal distribution of high/low priority projects, so that the whole group has a balanced workload. Many teams will employ individuals specializing in crystal production, data collection, and structure determination and refinement. However, the most efficient teams will have the flexibility and the expertise to adapt as needs change or demand in one particular area increases. This often requires a certain level of integration which was not always achieved in traditional crystallography laboratories. Most crystallography groups are relatively small and throughput demands high. So the overall success of an operation will also depend on the effective communication and interaction with groups outside of the crystallization process, such as protein engineers, for a stable and continuous protein supply.
In recent years, many crystallography laboratories have faced these challenges on top of those associated with the process of protein crystallization itself. Getting a protein to crystallize is often a rare event, but then establishing whether this is a viable crystal for structure determination presents further difficulties. Departments are under pressure to increase project turnaround to meet demand and to top it off the crystallization process is expensive and utilizes valuable protein reagents that are often time-consuming and costly to produce. These pressures have stimulated many crystallographers to look at implementing new technologies to aid their research efforts and maintain their position as financially independent and viable business units.
Technological advances have made new systems readily available to the market, and today there is a wide selection of equipment that can provide the protein crystallographer with the opportunity to improve throughput and increase experimental repeatability. However, the lab manager still faces many challenges. In order to successfully implement a new technology, they must feel confident that the equipment not only suits the task at hand but is better and more reliable than the manual alternative and can be installed with minimal disruption to workflow. Thus, one of the main objectives is to identify a technology that has enough application flexibility to provide reasonable future-proofing and is fully supported by the vendor. This sounds simple, but in a scientific discipline where precise and difficult-to-automate manual processes dominate, it is no mean feat. Additionally, along the road to technology acceptance, scientists must acknowledge that with automation generally comes a different way of working, and that is not always a bad thing, although occasional compromises must be made.
Technology evaluations should form part of a major discussion between the lab manager and the crystallography research team, to allow assumptions about current processes to be questioned, evaluated, and tested against the new methods. Often, changing the mindset of the laboratory staff is one if the most difficult challenges the laboratory manager faces when implementing new instrumentation and thus it has only really been in the last few years that crystallography automation has been more readily accepted. Previously, most crystallization related work was done more or less manually.
Many laboratory discussions are based around comparisons between person and machine. For example, if a lab predominantly wants to increase the speed of a process, i.e., liquid transfer, it is arguable that a person can be faster than a machine for some operations. However there are other considerations: machines do not tire, their accuracy does not dwindle, they do not pose the health and safety or RSI risks commonly experienced with manual methods, and they repeat experimental steps identically. Conversely, they cannot make informed decisions, and this can limit flexibility. It is important that the lab manager evaluates both the process and potential compromises to establish a more efficient and reliable experimental set-up. Common aims for the protein crystallography lab are to increase sample throughput, minimize expensive protein and reagent consumption, and increase crystallization success with the production of more stable crystals for data collection and structure determination. The screening of protein crystallography conditions is one area where protein consumption is high and experimental procedures are manually intensive. For these reasons, automation can provide many benefits.
FINDING AN AUTOMATED SOLUTION
The automation of crystallization screening presents two main areas of focus: liquid handling for plate set-up and imagers to inspect and record the results of the screens. The general consensus is that the streamlining of both presents significant financial benefits to the protein crystallography laboratory. Consequently, there are two aspects of this process that require consideration by the lab manager and his team. One is obviously time and the need to increase throughput without compromising accuracy. The other is to miniaturize the volumes of valuable protein and reagent solution used to improve cost efficiency. This often means that several strategies can/must be applied. For many larger laboratories, it is common to outsource reagent preparation to save time and improve screening reproducibility with batch consistency. Another option is to source a suitable liquid handler that can accurately pipette small volumes across a range of viscosities.
The selection and implementation of such instrumentation can often be a process of trial and error, as not all liquid handlers can maintain pipetting consistency across different plate types and aspiration volumes can differ with changing viscosities. However, there are instruments that can overcome these issues. A pipetting volume range of 50–1200 nL is ideal for screening crystallization conditions as it allows the set up of many conditions using minimal amounts of protein.
Crystallography groups that have incorporated such instruments into their laboratory workflow have noted many benefits, including a faster and more accurate set-up of drops compared to manual methods used previously, a vast reduction in protein consumption and cost for setting up screening plates, (in some cases nearly 10 times less per condition), screening of more conditions in 96-well SBS-format, and increased reproducibility of all liquid transfer steps. Many experiments, such as hanging drop set-ups, can be implemented in an easier way: mosquito simply places droplets of reservoir condition and protein on the plate seal in a mirror image of the plate, which, when inverted over the plate for sealing, locates each droplet over the correct well.
Additionally, harvesting crystals for successful data collection, using smaller amounts of protein has had real advantages upstream with protein production and purification groups, as some proteins may be particularly difficult to obtain, purify, or remain stable enough for storage. Thus, automation has radically improved the crystallization process. In fact today, the protein crystallography laboratory will frequently automate all or most liquid handling steps, with a large volume dispenser for reagent handling and nanoliter pipettors (often staged at two different temperatures, one at 4 °C and one between 20–24 °C) for screening or optimization set-up.
FIT FOR PURPOSE
There are of course other aspects of the crystallization screening and optimization process that need to be considered, namely the crystallography techniques implemented. Some laboratories have found that through the implementation of liquid handlers — more specifically a nanoliter pipettor — they have been able to improve and modify some crystallization protocols. Although sitting drop is very popular, especially with automated set-ups, there is still a great demand for the hanging drop technique. This is either due to the nature of the protein being used, such as membrane proteins which are notoriously difficult to crystallize, or to the fact that some individuals prefer to screen with hanging drops as this technique is commonly used for the subsequent optimization steps — thus transfer is easier and often more successful. As one of the more time-consuming techniques, automation of this method was not always available, and if it was, it was not necessarily reliable.
For a manager of a crystallography laboratory, whether it be large or small, automated liquid handlers are at the top of the list of “tools to have.” As far as automation goes they are not too expensive and, where labs have made the investment, they have benefited on a daily basis from increased experimental consistency. Liquid handlers also offer increased freedom for individuals to do the experiments they want or need to do. Additionally, some laboratories have opted to share both the investment and the use of such technology and can now successfully accommodate many users and optimized laboratory workflows.
In fact, the overall success of liquid handlers can also be measured by general acceptance. Previously, reluctant researchers who may have been less than confident with the idea of trusting a machine to do their work have been pleasantly surprised and adapted their workflow accordingly. This cautiousness is not uncommon and initially laboratory groups can shy away from automation. There is a misconception that automation means filling your laboratory with large integrated systems. Many crystallography facilities across the globe have proved that this is not the case and implemented liquid handlers as separate workstations. Although integration has its benefits in some laboratory set-ups, it can be too expensive and is often unnecessary for the crystallography laboratory. Many managers have found that separate workstations provide a greater flexibility, particularly for multiple user environments. Furthermore, an instrument can be easily removed from the operation for routine maintenance, with minimal disruption to operations. Of course, this is also where the relationship with the vendor, the quality of customer care and technical support, and the ease-of-use of the instrument come into their own.
ATTENTION TO THE FINER DETAILS
The lab manager is not only tasked with reviewing individual processes and how they can be best modified, but also looking at the bigger picture and how changes made to one system may affect another. One good example of this is the use of the microplate, a basic but essential consumable for the protein crystallography lab. Protein crystallography labs have strived to successfully incorporate image-capturing systems to accommodate the increased number of plates and crystallization screens generated. As a consequence they have also had to re-examine the microplates used. A shift from traditional hanging drops manually set-up on glass cover-slips to the use of automation to prepare sitting drops in plates covered with clear adhesive tape, is one such example that no doubt stimulated the adoption of the 96- well sitting drop SBS-format plate. To accompany the changes in crystallography methods and technological advances, consumable providers have recognized an increasing need to manufacture plates that are compatible with plate handling devices, liquid handlers, barcode readers, and imagers. Crystallographers now have the task of identifying the best one for their individual set-up.
The protein crystallography lab manager has faced many changes in the last decade. By readily adapting and reorganizing the laboratory infrastructure, and by using the recent technological advances to their advantage, it has been possible to significantly improve the efficiency of the crystallization process to meet the increasing demands. This has been particularly relevant for screening set-ups, where not only has automated instrumentation for liquid handling and image capture been successfully incorporated, but crystallographers have also worked to semi-automate their X-ray data collection. The automation of sample mounting and centering, has helped accommodate the increasing number of crystals coming through for testing using X-ray diffraction. The concept of automation is now more readily accepted and, where implemented so far, for example to miniaturize set-ups, it has benefited many labs, improving project turnaround time and cost efficiency. However, there is always room for improvement. With improved delivery often comes increased demand and the laboratory is expected to continuously improve its output. The demands placed upon the crystallography laboratory will not cease in coming years but instead will change, putting strain on new areas of the crystallization process and producing new bottlenecks. But as automation for crystallography increases in flexibility and sophistication, we can expect to see continued improvements to crystallography set-ups and in turn, throughput and project turnaround. Whether it is through further miniaturization and the incorporation of 384- well formats, development of software for intelligent design and auto-scoring algorithms, fully automated crystal mounting for Xray diffraction, or the use of in-house or table top synchrotons — only time will tell and the crystallography community waits with anticipation.
Joby Jenkins is a Product Manager for TTP LabTech.